Making Medical Prescription Automatic: The Case of Prescribing Therapies in Ankle Fracture Rehabilitation by Means of a Computer-Aided System

Automated decision support systems are computational tools that have been applied in clinical practice and have many benefits in processes, such as the diagnosis of diseases. Scientific and technological advancements in this area have led to the development of systems for other decision-making processes including medical prescriptions. In this study, we aimed to automate medical prescription making by designing a computer-aided system based on the experience of clinicians. In this case study, we automated the prescription of therapies for ankle fractures in a physical rehabilitation program. The database of the computer-aided prescription systems comprised a set of clinical records, from which the input variables were signs and symptoms related to ankle fracture rehabilitation, and the output variables represented rehabilitation therapies that may be prescribed by clinicians. The system was clinically validated, and its performance was quantified using confusion matrix metrics: 97.4% accuracy, 98.7% precision, 96.6% recall, 98.4% specificity, and 97.6% F-score. Therefore, the proposed system could be a useful tool in decision-making processes as prescription therapies that could contribute to the later motivation regarding traditional physical rehabilitation programs which is the optimization of resources for both patients and physical rehabilitation centers, while rehabilitation objectives are achieved.